Cache disk storage provides a means for storing data that is being migrated, for example, from a host system, to long term storage, for example, to magnetic tape. Data stored in cache disk storage is available for short term access by the migrating system, and is typically initially resident in cache disk storage, and then is premigrated, or copied to magnetic tape, so that the data is in two locations, the cache disk storage and the magnetic tape. Ultimately, the copy in cache disk storage is allocated to freespace and made available to be overwritten by new data, completing the migration.
Cache disk storage typically comprises a plurality of disk drives, and may take the form of one or more “drawers” or “blades” of a plurality of disk drives each, for example, as a RAID (Redundant Array of Independent Disks) or as a JBOD (Just a Bunch of Disks).
Eventually, the capacity of the cache disk storage needs to be increased, and this is typically accomplished by adding a drawer or blade to the cache disk storage, requiring a time consuming reconfiguration of the filesystem in accordance with the new capacity.
After the reconfiguration, the existing data is typically on the original drawer(s) or blade(s), and the added drawer or blade is empty. This means that new data is likely to be placed in the added drawer or blade, and not the original drawer(s) or blade(s), limiting throughput. Data is premigrated from the original drawer(s) or blade(s). When the added drawer or blade reaches capacity, new data is likely to be placed in the freespace of the original, and data is premigrated from the added drawer or blade, and a cyclic data transfer process continues, thereby affecting the cache disk storage data throughput.